What are some good systems for shutting down trolls in a co-op game?

News & General Discussion
1

I’m working on a small Versus Diablo style game where two teams of four race to get the last licks on “not-Diablo”.

Since it’s cooperative there are some design elements that are left wide open to trolling.

Namely, there will be a quick chat feature similar to Tribes or Smite where you can instantly send 65 different messages. The inventory will be a shared inventory so people can juggle loot fast.

Rage quit, chat spam, surrender spam, enemy feeding, inventory sabotage, and general toxicity seem the most likely candidates.

What have you seen that works or doesn’t work to stop trolls?

What can you see trolls doing that I may not have thought of?

Oh, and I should mention that I’m planing on making some characters that are specifically designed to troll the other team. Do you think that will keep them distracted?

2

The simplest is using report function.
With some automation and manual checking, you can filter/warn/punish out misbehaving players.

Simply collect reports per each player and when accumulates few of them in short period of time, validate them.

If you want go level up with that feature, you may want to implement replay system, for reviewing specific gameplays. Of course, replay should not be purely based on video recording. But deterministic inputs collection, to reduce number of stored data per replay. Then after provided time stamp, you can find exact moment what happened where, and can view action of every player.

In case of recorded video, you will be biased toward person, who recorded video. Sometimes may be clear, sometimes important parts may be cut out during editing.

2 Likes
3

I always liked the idea of moving trolls towards trolls.

4

That could be form of punishment, sure.
It also depends on level of misbehaving.
Weather to ban or not.
Or simply reallocating to other group of bad behaving players.
But how you give them an opportunity to come back?
You probably want some scoring / warring level system, which can decay over the time?

5

Convenience > everything else.

If it’s more convenient to cooperate than to troll, that probably make the biggest impact.

Specific solution, I dunno what could be offered without being knee-deep in the project. I wouldn’t sweat over it too hard though. I’ve never played any MP game without it’s trolls. If AAA can’t solve the issues with expensive management departments just to handle community behavior, I wouldn’t waste much time as one man show worrying over it. It’s not the thing that makes or breaks a game, and people are capable of policing themselves + forming their own groups.

1 Like
6

I hope you get enough players to have this problem.

7 Likes
7

I’ll elaborate a little more – and this is kind of beyond the original question but it gives more context to my answer. So, when you are making a game one thing you probably already doing is you ask questions about time. “How long will it take to code this feature? Is the time investment worth the final outcome?”

You got to do the same thing with decision making. “How long will it take me to figure out a good solution. Would the most optimal solution be worth that time? Would a less-than-optimal solution be worth that time?” “What is the likelihood of me reaching an optimal solution? Or any solution?”

So you got to prioritize the time you use to make decisions. Also consider that just like time, brain power is a finite resource and you got to portion it to the most bang-for-the-buck places.

Now, I’m not a professional developer so excuse me if I’m speaking a bit out of my ass here, but it seems to me a difference in amateur developer versus professional is that professional plans around the capabilities of the team divided by time, while amateur only plans based on their desires.

Similarly, in the military you don’t plan for best case scenario. You plan for worst case scenario every time, and whatever plan you make you spend every second you got trying to imagine every way it can fail. Most of the failures are going to be human related, so that is where you put most your time into.

Before you make any plan, you need a plan for how you are going to spend your time planning, right? Sounds silly but if you don’t have lots of experience in this, chances are you waste your time planning and then you go off half-cocked.

1 Like
8

I think topic is still very valid. Even if I or OP won’t be using such mecanics, someone else may find it perhaps interesting.

While having spending time on developing such feature is valid concern, or whether user base is of enough size, I would like to hear more opinions, replaying to OP question, even if is purely hypothetical.

Or maybe rephrase OP question. How to handle toxicity in game chat?

9

What works:

  • disable chatting: can’t spew toxicity if you can’t talk
  • put a report option, if the player is reported enough and has low activity, ban them temporarily
  • create an AI to track player activity and idling
  • create MVP rewards that cannot be gotten otherwise. if you make them compete to be good, it is hard to want to troll
  • make clear objectives for players to cooperate on based on proximity

What doesn’t work

  • report option based on number of reports, trolls abuse it to ban less experienced players
  • banning quitters, sometimes you have to go
  • vote-kick with reasons, often misused.
3 Likes
10

I think match making based on ranks, could help reduce that issue.

However, there is occasional quitting and often quitting in team based games. That behavior can be detected easily.
Specially when quits happen, during team starting loosing, or at beginning of the match.

Even good (too good) players can be voted by others to quit. Agree. I think only checking replay after game can validate reason for kick.

Zero-K handles such issues, with in game voting system, where most players must vote, to pass the vote, and of course, +50% of players votes in particular match must be casted, which agree with given vote, to pass resolution.
And resolutions may muting, kicking, resigning etc. System works even in lobby before game, to vote for map change and teams shuffle. Voting system is separate from reporting and banning.

The issue sometimes is with that mechanics, when at least 50% players in game / lobby don’t vote. Resolution can not be passed then. Also, sometimes vote system can be abused, when many new players is present. Meaning, such players vote yes anyway, as don’t understand fully the reason. Leading to abusive cast resolutions. Yet I found, such cases are relatively rare. Eventually naughty player is kicked.

Not sure, if similar system will work, with none-team based multiplier sessions. Perhaps some voting resolutions could be adapted.

11

I’m not a huge fan of banning players.

I would much rather shame or quarantine them.

I could see doing something such as keeping track of what they say and if they do “You rock! Cancel that!” again and again I would give them a giant head of a crying baby for their player. If they kept it up, then I would start them in matches muted on other players. Keep it up more and they get sent to the shadow realm of troll accounts. And finally, start putting them in games with bots and maybe even start rigging the game for them to lose until they quit.

Then the whole prestige value could also determine your queue in line, or things like that.

1 Like
12

I have very little direct experience with this topic, but I’ve always felt that a reputation system ought to work if the algorithms behind it are sufficiently clever.

So for example suppose you have a convenient way for a player to give another player a thumbs-up (or thanks or tip o’ the hat) to other players who have been helpful or role-played well. And you have a somewhat less convenient way to give a thumbs-down (report) players who are being rude or abusive. But then you don’t just act on the raw numbers — you have some clever algorithm that looks at the patterns of thumbs-up and thumbs-down to identify the trolls.

To me, it doesn’t seem that hard; trolls will be players that both give and receive a lot of thumbs-downs. Or so I would expect. And then you can start applying those consequences like you described above. Personally, I like the idea of putting them in matches with bots, especially if they can’t tell they’re bots.

2 Likes
13

I also like thumbs up and thumbs down, but I’d really like to add a once a day super thumbs up and super thumbs down.

That way if someone gets multiple super thumbs down in one match you know they’re a problem

2 Likes
14

You’d think that but in many games this is either abused by players that are friends or against bad players. You cannot get better through this system. And then it becomes an entry issue, new players will not adhere to such a game.
He mentions trolls but describes them with a wide brush: quitters are trolls. How? I get the people that are toxic, but removing chat usually works.

15

You’ve obviously never played smite. Despite having no insult function, people still manage to ruin games by spamming “you rock! Cancel that!” Then sit in the fountain spamming a surrender vote. And failing that throwing themselves at the enemy then quitting.

Once teams are 3 v 4 there’s almost no coming back.

I once pulled out a win 2v5 because I stacked out my toon to sit in the tower and the other player took it. Which was amazing. But definitely the exception and not the rule.

16

Oh I have, and I know all about this. This is as old as AOE 2 days.

If you want to build around quitters: make an AI possess the player on a difficulty based on current score.
Surrender votes should not exist. Human nature has it that people give up quite easily. In a 1v1, sure. In a team game, heck no!

17

This just reminded me RTS Zerok-K case. Unfortunately, following will not work in other type of game. Games are typically 2 to 32 players. There are often cases, multiple players quit during battle. So how the problem is tackled, of abandoned units in this RTS, that all units of exiting player are given to the most experienced player (ranked). Same applies, when player become idle for more than few min or so. But when player become active again, or logging back to the game (i.e. after occasionally power cut/game crash), player receive their units back, other than new built units.
Really cool feature, making such issue at most mild annoying. Unless different players drop abandoning their bases in different parts of map. Is becoming hard to manage such spread forces in large maps and dynamic games.

18

Sounds like a great way for a group of friends to try to ban innocent players they dislike.

I liked @JoeStrout 's idea about weight based on how often they report.

Someone who gives a thumbs down to everyone, every game, should be dismissed as a troll themselves. While someone who rarely or never reports, would likely only do so if it is really really bad.

I personally would have two separate ways to report. One being for chat, which is easily accessible, and does nothing, and a more inconvenient report for negative action (feeding, etc.) Sticks & stones may break your bones, but words wont hurt your gaming.

I also am of the camp that you should never under any circumstances except Cheating, ban anyone for anything. It is not only contrary to good business but IMO a violation of Consumer’s Rights.

Instead, I think a big problem lies in the designs of many games. MOBA games are among the most toxic by such an enormous margin. Even their tournament players get banned for toxicity. It is that bad. However the game’s design encourages this toxicity. Not only does it not discourage it, it encourages it. The game is purposefully designed in a way which makes even good people toxic.

That is in how unfriendly MOBA games are. You are punished for leaving a match. You are banned for it. You are punished by being unable to surrender early even if youre being trolled.

This alone causes a huge amount of toxicity. If youre no longer having fun, the entire purpose of a video game, and the developer locks you in to waste your life for 20-40 minutes, a good way to get it back is to troll, piss everyone off, and thus have fun. Those players arent even necessarily bad. The game is bad for locking them in by force.

So long before you come up with ideas of how to creatively punish players, you should consider fixing your design problems instead.

I guarantee a huge amount of toxicity in MOBA games would be resolved simply by respecting the player’s time more. It would also raise the medium age, as adults are much less likely to waste their lives away having no fun for 20-60 minutes than jobless teenagers.

I stopped playing MOBA games myself many years ago simply due to how little the developers respected my time. I have better things to do than wait 40 minutes for my team to lose simply bc 1 second in someone disconnected.

I also wont lie. I know how to manipulate most multiplayer systems. I can ban anyone I want in a few MMO’s just by running a few virtual machines. I have never done it, but I know those who have. Pissing off the wrong guild can also get you banned very quickly. Quantity Reporting (instead of quality) and Automated Bans are ripe for abuse by people smart enough to use multiple accounts or have alot of friends/guildmates.

In some cases, you’re actually introducing more trolling via report features than you are preventing. This is especially true with automation, but even manual bans from human employees are still easily biased by quantity of reportiny by trolls/guilds. This is why IMO bans should only be reserved for cheaters. In many games, you can read pages upon pages of angry customers of MMOs who were unfairly banned via automation or paid employees having a bad day.

3 Likes
19

Let players form their own 4 person teams with their friends, and disable random team members even if they only have 2 or 3 on their team. This isolates cooperative trolling only to teams picking up randoms, providing players their own way to eliminate the problem. Try to match teams of 2 or 3 with other teams of 2 or 3 when available.

Add a report function as already mentioned. I’d calculate a normal reports to total games ratio and review players who are higher than the normal ratio.

I too like the idea of isolating trolls, or even cheaters, to play only with themselves rather than banning them outright.

2 Likes
20

Weighting reports based on the volume of reporting they do is nothing short of brilliant. Thanks Joe.

This whole using the report function as a troll thing is a very valid point.

It seems like a better solution than nuking it all would be to simply count all players in a match who are friends/joined together as one report.

2 Likes